microsoft/Phi-4-reasoning-vision-15B
Full fine-tune (fewer layers) vs microsoft/Phi-4-reasoning. microsoft/Phi-4-reasoning matches part of microsoft/Phi-4-reasoning-vision-15B (243 of its 695 tensors; the rest of microsoft/Phi-4-reasoning-vision-15B has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 40 of 40 layers.
Tensors changed: Embeddings 1/1, Norms 81/81, MLP 80/80, Attention 80/80, lm_head 1/1
4 derivatives checked
- DevQuasar/microsoft.Phi-4-reasoning-vision-15B-GGUF: Quantization (fewer layers)
- dranger003/Phi-4-reasoning-vision-15B-GGUF: Quantization (fewer layers)
- jamesburton/Phi-4-reasoning-vision-15B-GGUF: Same weights, converted (fewer layers)
- kingjones777/Phi-4-Reasoning-Vision-15B-ROCmFP4-GGUF: Quantization (fewer layers)
4 other repos hold byte-for-byte identical weight files.